Dan's Data letters #125
Publication date: 13 September 2004.Last modified 03-Dec-2011.
The flipped-bit blues
I recently purchased a cheap hard drive enclosure for $AU85. I looked it up on the Net and found that it was made by Zynet. It has only a USB2.0 link, no FireWire, and I'm using it with my WinXP PC.
The problem I have is that when I transfer my gigantic DivX AVIs across, the file does not transfer across properly. By that I mean the data is stuffed up every now and again. File size is identical, but the files have glitches in them. I thought was a problem occurring only in transferring large files, but even small files are corrupt now and then.
Is this problem unique to cheap hard drive enclosures? Or is there a problem with USB 2.0? I'm inclined to believe that my enclosure is crap and I'm gonna get it swapped for something else or get my money back. I'd appreciate any info you have on this kind of problem.
Paul
Answer:
There are a few places where the corruption could be occurring.
Yes, it could be the enclosure's bridge circuitry; probably not something as simple as a loose connection or bad cable, or your OS would notice the problem.
It could also be a lousy USB controller on your motherboard. NForce2 boards, in particular, are famous for USB problems, though data corruption is not usually among those problems. Sometimes USB controller problems can be fixed with BIOS upgrades or other tweaks, but most people take the simpler way out and just drop some dollars on a PCI USB controller card.
It could also be a problem with the drive itself. Dying hard drives often cause data corruption, and corruption will be often be noticeable in giant video files before it shows up somewhere else, partly simply because they're so enormous (and so more likely to be hit by corruption that happens every X kilobytes), and partly because many video file formats are "fragile", and can't tolerate any junk data Bad bits in the middle of an MP3 just make a brief nasty noise; bad bits in the middle of many kinds of AVI can crash the player. This problem you can probably rule out by trying the drive on a regular IDE cable.
(Also, it could just be that you're neglecting to "Zest Your Need, Enjoy Together!")
Fearing the reaper
What, exactly, can cause plain electronics to die? I'm using an old Slot 1 Celeron 366 box as an ADSL router - it runs Coyote Linux (your suggestion - thanks muchly), and features a floppy drive and two network cards. It has exactly one fan, in the power supply, and I'm tempted to pull that out and do something else with it (an idle 366 makes little heat). I'm wondering about how long I could reasonably expect it to last... without dust (practically none already, and there's almost no airflow), or static problems (not likely to crack open the case again), or disk drives (the floppy drive averages maybe a couple of minutes a year of actual use), what could go wrong, if anything?
Tim
Answer:
I think the primary killer of non-mistreated CPUs (no static damage, no ferocious overclocking or long-term overheating)
is electromigration - the slow process by which electrons
push metal atoms down conductors, including the interconnects between the transistors in CPUs.
There are plenty of other non-moving parts that can die in a PC, though. Electrolytic capacitors will dry out eventually, in particular. A motherboard can survive several dry electros that've moved way off their proper values, since they're mainly just power supply smoothing caps (this is why you can knock some caps clean off a motherboard and have it seem to work fine), but sooner or later the camel's back will break. Various other components can, similarly, fail, leaving you with a dead or severely flaky computer which may still have an A-OK processor and RAM - but, by then, every motherboard that can use that OK CPU and RAM will be just as old, so you might as well upgrade to something only seven generations behind the cutting edge rather than stick with your 12-generations-behind hardware.
How many years a simple old machine like yours will last is hard to say; there are still IBM XTs and ATs chugging along out there somewhere, after all. It could definitely only be half way through its life today.
Or a big cap in the PSU could explode tomorrow.
Regarding losing the fan: If the CPU's currently passively cooled, you may be surprised how hot the machine manages to get when you remove the feeble breeze from the PSU fan. A good compromise is to hack a big hole in the case, mount a 120mm 12V fan on it, and run that fan at reduced voltage - 5V, if it'll start up at that voltage. Very quiet, very long lived, and as much or more air flow as the PSU fan is probably delivering.
You forgot "Super Nashwan"
Why do they sell ordinary batteries using buzzwords like "Heavy Duty", "Super Heavy Duty", "Ultra", "Max", etc, while they sell rechargeables with much more informative ratings, such as their capacity in milliamp-hours?
Not only do they seem to not market batteries with these ratings, they almost always fail to mention them on the packet. Sure, sometimes you can find the data sheets on the Net, but it doesn't help you much when you're sitting in the battery aisle wondering how you should rank "Max", "Ultra", "Whiz Bang", "Supreme" and "Super Heavy Duty Gold". I guess it's fairly safe to assume that a known brand is better, but how good an indicator is price when comparing the big brands?
Iceman
Answer:
The main consumer rechargeable technologies - NiCd and NiMH - both have low internal resistance and high current delivery
capacity. What that means is that an accurately specified 1000 milliamp-hour (mAh) cell will, quite reliably, give
you ten hours of service at 100 milliamps (mA), or one hour of service at 1 amp (A). Rechargeable cells also have
rapid self-discharge - they go flat very fast just sitting there, compared with non-rechargeables - so their effective
capacity for very low loads (wall clock, calculator, remote control...) is much lower than you'd expect, but for substantial
loads their charge life can be worked out very easily.
[UPDATE: This page is a few years old, now, and it's now easy to buy "low self-discharge" NiMH cells, which have lower absolute capacity but hold their charge much better than the standard kind.]
Carbon-zinc ("dry") and alkaline cells, though, have high internal resistance and low current delivery capacity; if you ask for a lot of current from them, they heat up and go flat much faster than you'd expect. A carbon-zinc cell with 1000mAh capacity into a 10mA load may have only 600mAh capacity at 100mA, and 250mAh or less at one amp. Alkaline's better for high current, and lithium is a bit better again (though not so much better that lithium AAs are worth the extra money, if you don't need their monstrous shelf life and/or good performance below freezing), but all of these technologies (plus the not-too-useful rechargeable alkaline) really don't want to be running flashlights, cameras or boom boxes, if they've got a choice.
For this reason, battery companies do publish quite exhaustive stats on their products - as you say, on sites like data.energizer.com and duracell.com/oem - but they don't quote them in the ads, because there's no way to quickly compare capacities when you don't know what load the batteries are going to face.
Personally, I'm annoyed by the alkaline battery advertisements that say that "Brand X's alkaline cells last three
times as long*!"
* as Brand Y's Super Heavy Duty non-alkaline cells.
Well, whoopee. Brand Y's alkalines are better than Brand X's carbon-zincs, too.
Price is not a very good indicator of performance. I don't recommend you buy off-brand carbon-zinc batteries, as all dry cells consume their casing as part of their basic chemistry and care therefore needs to be taken to avoid nasty leaks; dirt-cheap dry cells can ruin your gear. But off-brand alkalines these days are generally excellent value for money, especially if you buy a pack of 24 or something. Yes, they may have a bit less capacity than the equivalent Duracell or Energizer (and will be a bit further again behind the premium "Ultra" or "E2" variants), but the milliamp-hours-per-dollar ratio is much better for yum-cha alkalines, as long as they haven't been sitting on a dock somewhere for five years before you get to buy them.
Waiting for the waterbed version
My girlfriend sent me a link to the Quantum Sleeper. I've no idea how she found it. I was wondering if you had heard of it. It would appear to be some sort of bed/safe room in one. Hey, maybe you can get them to send you one for review!
Grey
Answer:
The Quantum Sleeper had its 15 minutes of fame on the gadget
blogs a little while ago.
My first thought when I saw it, of course, was how best to kill the Lifestyles Of The Rich And Paranoid twit hiding inside it. Smiling and waving at him while you splash gasoline all over the bedroom could be an idea - the Quantum Sleeper's air scrubber would prevent rapid asphyxiation, and polycarbonate's pretty fire-resistant, but it'd still be an experiment well worth doing. Setting up a sparkler bomb on the top of the clear portion could be fun, too.
Polycarbonate is, of course, quite easy to drill through. If it's a polycarbonate/glass laminate like most bank-teller windows then you might spoil your handy-dandy brad-point drill bit in the process of bulling through the stuff, but the guy inside the Quantum Sleeper presumably won't have any way to stop you plugging in whatever power tools you like, so you won't even have to bother with an expensive heavy-duty cordless drill.
And the expression on his face as you start squirting the nitric acid through the hole will make it all worthwhile.
From the Don't Use Your Mum's Stove department
I recently read an article about Abbie Hoffman which had a recipe in it that said mixing and slowly cooking potassium nitrate and sugar into a gel, and letting that gel cool into a solid, produces an excellent smoke producing device when lit with a wick, flame, match, etc.
Does this really work? What are proportions if it does? And is it easy to purchase potassium nitrate?
Matthew
Answer:
Yes, it does work. And, because it's not an explosive, you're not running a terrible risk of blowing yourself through
the kitchen door in a red fog in the course of making it. The mixture does, of course,
really want to catch fire during
the melting process, and if it does it'll burn with great enthusiasm (so melt it on an electric stove...).
As long as you're wearing eye protection, though, the worst it's likely to be able to do to you is give you some stylish
scars, as it sets fire to your kitchen. Hurrah.
I do not, of course, encourage you to actually do this, and if you do do it, destroy your house and get full-thickness burns to 80% of your body as a result, please don't get your heirs to sue me. But there are many Web pages on the subject of making this stuff. Which, as a reader's now pointed out to me, can make a better-than-decent rocket fuel, as well as being a smoke compound.
Whether you can buy potassium nitrate (KNO3) depends on where you are. It's saltpetre - the key ingredient in the manufacture of good old black gunpowder.
(As every bomb-crazed 13-year-old boy knows, sulfur's not hard to find, and you can make your own charcoal; KNO3 can be made at home, too, but you've got to start with rather a lot of manure and/or fermented urine, which is beyond the commitment threshold of most amateur anarchists.)
Because of this, purchasing saltpetre is, generally, difficult in countries where purchasing gunpowder is also difficult.
After 9/11, it has of course not gotten easier to buy pyrotechnic ingredients in most Western nations, and if you're dicking around with this stuff and the cops come to call for whatever reason, you can count on news reports of your arrest containing terms like "cache" and "bomb-making equipment".
If all you want is a bunch of smoke, I recommend you look into water based smoke fluid, for commercial smoke machines and certain toys, instead. You can get a nasty non-repairable smoke machine quite cheaply (eBay's a good source of 'em), but you don't actually need one to turn the water/glycol/herbs-and-spices fluid, which you can buy from various stagecraft and lighting shops, into amazing amounts of smoke. All you need to do is dribble the fluid onto something rather hot, like an electric stove hotplate. (Amusement can also be gained by spraying the fluid up exhaust pipes or onto engine blocks.)
The fluid's quite expensive per litre, but it'll never go bad, a little goes a long way, and it's way safer than pyrotechnic smoke compounds of all kinds; you can use glycol smoke fluid to put fires out. Glycol fluid smoke is also safe to breathe; the only way it'll hurt you is if you get your flesh in the way of the vapour before it's cooled enough to condense into smoke. (Some people claim to have some sort of allergic reaction to glycol smoke fluid, but I think many of them have mistaken oil-cracker smoke for water-based smoke, or just reckon that smoke that dense must be bad to breathe, and hence have a psychosomatic reaction to it.)
As long as you don't put your hand an inch in front of a smoke machine nozzle, you'll be quite safe.
Enough, already
Black Hole Cables look like something you might be interested in. It's your standard overpriced-cables scam, but with a twist - they intend these for your PC. Hundred-dollar power cables can now not only make your stereo sound better, they can also make your PC faster. Or something. To their credit, Black Hole isn't claiming anything too crazy about their cables. The $40 Ethernet cables are pretty funny though.
There should be a joke in here about throwing money into a black hole, but that would be too easy. Here's a review by somebody that should know better.
Ryan
Answer:
Ryan was just the first of several people to point out Black Hole Cables to me. They're not nearly as expensive as
the higher end audio cables, but yes, the basic concept is still ridiculous, particularly when you note that the "Pro-Grade"
cables used by actual computing "Pros" are, as is the case with audio, entirely ordinary voodoo-free products. In
many cases, they're the same as the ones they sell at your local computer/electronics store.
Toughening up unshielded twisted pair Ethernet cables is pretty pointless, too. Where do those cables fail? Well, often nowhere if they're the "booted" kind that don't have the fragile RJ45 retention clip hanging out in the breeze, but otherwise, they fail at the RJ45 connectors. The Black Hole RJ45s are billed as "superior", and have a screw-on strain-relief collar behind them - but if you're a "pro", you should be setting up your network cabling so that nothing's tugging on a plugged-in cable anyway. These cables need some extra strain relief, to support the relatively massive weight of their pointless metal sheathing, but the plugs themselves look to be the same weak clear plastic as everyone else's.
As for the Black Hole power cables... where do $1 IEC leads fail? Uh... actually, I don't think I've ever seen one fail in any way. Not the cable, not the plugs, not the contacts; really old ones may get loose at the IEC-plug end, and I'm sure people have managed to mangle them by running chairs over them repeatedly or slamming them in doors, but I never have. But, again, at least the Black Hole one is really heavy; as the overclockers.com reviewer noticed, you've got to support it to make sure it doesn't unplug itself under its own weight.
I also like the proofing errors in the Black Hole PDFs - this one, for instance, currently says that a "patient" has been applied for, among other things.
There must be an easier way
The vehicle: A motorcycle powered by 24 chainsaw engines.
The question: Why???
Kath
Answer:
The Dolmette site makes it pretty clear that this extraordinary contraption
is just a promotional novelty piece for Dolmar. The end specs of the ludicrous engine aren't that incredible, anyway;
for all the chassis room it takes up, it only gives aggregate power and torque similar to that of somewhat-stock-looking
four stroke drag bikes with similar
total displacement,
and it would of course be totally annihilated by the wheelie-bar
brigade.
I bet it was a much more entertaining engineering project than any of the chainsaw company's nerds expected to be working on, though.
On the subject of big silly things with lots of little engines, one of the world's foremost examples bit the dust the other day.
